aptos-imageresizing-resnet-101-kappa-optimization.ipynb

{
 "cells": [
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {
    "_cell_guid": "b1076dfc-b9ad-4769-8c92-a6c4dae69d19",
    "_uuid": "8f2839f25d086af736a60e9eeb907d3b93b6e0e5"
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "['resnet101', 'resnet50', 'aptos2019-blindness-detection']\n"
     ]
    }
   ],
   "source": [
    "# This Python 3 environment comes with many helpful analytics libraries installed\n",
    "# It is defined by the kaggle/python docker image: https://github.com/kaggle/docker-python\n",
    "# For example, here's several helpful packages to load in \n",
    "\n",
    "import numpy as np # linear algebra\n",
    "import pandas as pd # data processing, CSV file I/O (e.g. pd.read_csv)\n",
    "\n",
    "# Input data files are available in the \"../input/\" directory.\n",
    "# For example, running this (by clicking run or pressing Shift+Enter) will list the files in the input directory\n",
    "\n",
    "import os\n",
    "print(os.listdir(\"../input\"))\n",
    "from shutil import copyfile\n",
    "from fastai.vision import *\n",
    "import cv2\n",
    "from sklearn.metrics import cohen_kappa_score\n",
    "import scipy as sp\n",
    "\n",
    "\n",
    "# Any results you write to the current directory are saved as output."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {},
   "outputs": [],
   "source": [
    "def random_seed(seed_value, use_cuda):\n",
    "    np.random.seed(seed_value) # cpu vars\n",
    "    torch.manual_seed(seed_value) # cpu  vars\n",
    "    random.seed(seed_value) # Python\n",
    "    if use_cuda: \n",
    "        torch.cuda.manual_seed(seed_value)\n",
    "        torch.cuda.manual_seed_all(seed_value) # gpu vars\n",
    "        torch.backends.cudnn.deterministic = True  #needed\n",
    "        torch.backends.cudnn.benchmark = False"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {},
   "outputs": [],
   "source": [
    "random_seed(101, True)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {
    "_cell_guid": "79c7e3d0-c299-4dcb-8224-4455121ee9b0",
    "_uuid": "d629ff2d2480ee46fbb7e2d37f6b5fab8052498a"
   },
   "outputs": [],
   "source": [
    "Path('/tmp/.cache/torch/checkpoints/').mkdir(exist_ok=True, parents=True)\n",
    "!cp '../input/resnet101/resnet101.pth' '/tmp/.cache/torch/checkpoints/resnet101-5d3b4d8f.pth'"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {},
   "outputs": [],
   "source": [
    "# from fastai_slack import read_webhook_url, SlackCallback\n",
    "# hook = 'https://hooks.slack.com/services/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'\n",
    "# slack_cb = SlackCallback('surbhi', hook, frequency=1)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "metadata": {},
   "outputs": [],
   "source": [
    "df_train = pd.read_csv('../input/aptos2019-blindness-detection/train.csv')\n",
    "df_test = pd.read_csv('../input/aptos2019-blindness-detection/sample_submission.csv')\n",
    "\n",
    "x_train = df_train['id_code']\n",
    "y_train = df_train['diagnosis']"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "metadata": {},
   "outputs": [],
   "source": [
    "# SIZE=224\n",
    "\n",
    "# train_df=pd.read_csv(PATH+'/train.csv')\n",
    "# test_df=pd.read_csv(PATH+'/sample_submission.csv'"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "metadata": {},
   "outputs": [],
   "source": [
    "PATH = \"../input/aptos2019-blindness-detection/\"\n",
    "\n",
    "train = ImageList.from_df(df_train, path=PATH, cols='id_code', folder=\"train_images\", suffix='.png')\n",
    "test = ImageList.from_df(df_test, path=PATH, cols='id_code', folder=\"test_images\", suffix='.png')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "0    1805\n",
       "2     999\n",
       "1     370\n",
       "4     295\n",
       "3     193\n",
       "Name: diagnosis, dtype: int64"
      ]
     },
     "execution_count": 9,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "y_train.value_counts()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "metadata": {},
   "outputs": [],
   "source": [
    "def crop_image(img,tol=7):        \n",
    "    mask = img>tol\n",
    "    return img[np.ix_(mask.any(1),mask.any(0))]\n",
    "\n",
    "def open_aptos2019_image(fn, convert_mode, after_open)->Image:\n",
    "    image = cv2.imread(fn)\n",
    "    image = cv2.cvtColor(image, cv2.COLOR_BGR2GRAY)\n",
    "    image = crop_image(image)\n",
    "    image = cv2.resize(image, (224, 224))\n",
    "    image = cv2.addWeighted(image, 4, cv2.GaussianBlur(image , (0,0) , 128/10) ,-4 ,128)\n",
    "    return Image(pil2tensor(image, np.float32).div_(255))\n",
    "\n",
    "vision.data.open_image = open_aptos2019_image\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "metadata": {},
   "outputs": [],
   "source": [
    "#np.random.seed(42)\n",
    "tfms = get_transforms(do_flip=True,flip_vert=True,max_rotate=360,max_warp=0,max_zoom=1.1,max_lighting=0.1,p_lighting=0.5)\n",
    "data = (\n",
    "    train.split_by_rand_pct(0.2)\n",
    "    .label_from_df(cols='diagnosis', label_cls=FloatList)\n",
    "    .add_test(test)\n",
    "    .transform(tfms, size=224)\n",
    "    .databunch(path=Path('.'), bs=32).normalize(imagenet_stats)\n",
    ")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 12,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 504x432 with 9 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "data.show_batch(rows=3, figsize=(7,6))\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 13,
   "metadata": {},
   "outputs": [],
   "source": [
    "def qk(y_pred, y):\n",
    "    #y_pred = torch.argmax(y_pred, 1)\n",
    "    return torch.tensor(cohen_kappa_score(torch.round(y_pred.float()), y, weights='quadratic'),device='cuda:0')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 14,
   "metadata": {},
   "outputs": [],
   "source": [
    "learn = cnn_learner(data, models.resnet101, metrics=[qk], model_dir=\".\", callback_fns=ShowGraph)\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": 15,
   "metadata": {},
   "outputs": [],
   "source": [
    "# learn.lr_find()\n",
    "# learn.recorder.plot()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 16,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: left;\">\n",
       "      <th>epoch</th>\n",
       "      <th>train_loss</th>\n",
       "      <th>valid_loss</th>\n",
       "      <th>qk</th>\n",
       "      <th>time</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <td>0</td>\n",
       "      <td>1.998085</td>\n",
       "      <td>0.961344</td>\n",
       "      <td>0.567835</td>\n",
       "      <td>04:53</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>1</td>\n",
       "      <td>0.929754</td>\n",
       "      <td>0.563048</td>\n",
       "      <td>0.779281</td>\n",
       "      <td>04:45</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>2</td>\n",
       "      <td>0.703320</td>\n",
       "      <td>0.412456</td>\n",
       "      <td>0.858596</td>\n",
       "      <td>04:42</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>3</td>\n",
       "      <td>0.616097</td>\n",
       "      <td>0.437380</td>\n",
       "      <td>0.812617</td>\n",
       "      <td>04:41</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>4</td>\n",
       "      <td>0.523823</td>\n",
       "      <td>0.363184</td>\n",
       "      <td>0.866365</td>\n",
       "      <td>04:20</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>5</td>\n",
       "      <td>0.507207</td>\n",
       "      <td>0.393462</td>\n",
       "      <td>0.841524</td>\n",
       "      <td>04:23</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>6</td>\n",
       "      <td>0.478148</td>\n",
       "      <td>0.340016</td>\n",
       "      <td>0.868926</td>\n",
       "      <td>04:40</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>7</td>\n",
       "      <td>0.393011</td>\n",
       "      <td>0.321175</td>\n",
       "      <td>0.867015</td>\n",
       "      <td>04:42</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>8</td>\n",
       "      <td>0.363363</td>\n",
       "      <td>0.292670</td>\n",
       "      <td>0.881872</td>\n",
       "      <td>04:42</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>9</td>\n",
       "      <td>0.320692</td>\n",
       "      <td>0.277182</td>\n",
       "      <td>0.896068</td>\n",
       "      <td>04:43</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>"
      ],
      "text/plain": [
       "<IPython.core.display.HTML object>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Better model found at epoch 0 with qk value: 0.5678347945213318.\n",
      "Better model found at epoch 1 with qk value: 0.779280960559845.\n",
      "Better model found at epoch 2 with qk value: 0.8585955500602722.\n",
      "Better model found at epoch 4 with qk value: 0.8663649559020996.\n",
      "Better model found at epoch 6 with qk value: 0.8689256906509399.\n",
      "Better model found at epoch 8 with qk value: 0.881872296333313.\n",
      "Better model found at epoch 9 with qk value: 0.8960679769515991.\n"
     ]
    }
   ],
   "source": [
    "learn.fit_one_cycle(10, 1e-2, callbacks=[callbacks.SaveModelCallback(learn,monitor='qk',mode='max', name='best_model')])\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 17,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "Learner(data=ImageDataBunch;\n",
       "\n",
       "Train: LabelList (2930 items)\n",
       "x: ImageList\n",
       "Image (1, 224, 224),Image (1, 224, 224),Image (1, 224, 224),Image (1, 224, 224),Image (1, 224, 224)\n",
       "y: FloatList\n",
       "2.0,4.0,1.0,0.0,0.0\n",
       "Path: ../input/aptos2019-blindness-detection;\n",
       "\n",
       "Valid: LabelList (732 items)\n",
       "x: ImageList\n",
       "Image (1, 224, 224),Image (1, 224, 224),Image (1, 224, 224),Image (1, 224, 224),Image (1, 224, 224)\n",
       "y: FloatList\n",
       "0.0,2.0,0.0,0.0,0.0\n",
       "Path: ../input/aptos2019-blindness-detection;\n",
       "\n",
       "Test: LabelList (1928 items)\n",
       "x: ImageList\n",
       "Image (1, 224, 224),Image (1, 224, 224),Image (1, 224, 224),Image (1, 224, 224),Image (1, 224, 224)\n",
       "y: EmptyLabelList\n",
       ",,,,\n",
       "Path: ../input/aptos2019-blindness-detection, model=Sequential(\n",
       "  (0): Sequential(\n",
       "    (0): Conv2d(3, 64, kernel_size=(7, 7), stride=(2, 2), padding=(3, 3), bias=False)\n",
       "    (1): BatchNorm2d(64, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
       "    (2): ReLU(inplace)\n",
       "    (3): MaxPool2d(kernel_size=3, stride=2, padding=1, dilation=1, ceil_mode=False)\n",
       "    (4): Sequential(\n",
       "      (0): Bottleneck(\n",
       "        (conv1): Conv2d(64, 64, kernel_size=(1, 1), stride=(1, 1), bias=False)\n",
       "        (bn1): BatchNorm2d(64, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
       "        (conv2): Conv2d(64, 64, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)\n",
       "        (bn2): BatchNorm2d(64, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
       "        (conv3): Conv2d(64, 256, kernel_size=(1, 1), stride=(1, 1), bias=False)\n",
       "        (bn3): BatchNorm2d(256, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
       "        (relu): ReLU(inplace)\n",
       "        (downsample): Sequential(\n",
       "          (0): Conv2d(64, 256, kernel_size=(1, 1), stride=(1, 1), bias=False)\n",
       "          (1): BatchNorm2d(256, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
       "        )\n",
       "      )\n",
       "      (1): Bottleneck(\n",
       "        (conv1): Conv2d(256, 64, kernel_size=(1, 1), stride=(1, 1), bias=False)\n",
       "        (bn1): BatchNorm2d(64, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
       "        (conv2): Conv2d(64, 64, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)\n",
       "        (bn2): BatchNorm2d(64, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
       "        (conv3): Conv2d(64, 256, kernel_size=(1, 1), stride=(1, 1), bias=False)\n",
       "        (bn3): BatchNorm2d(256, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
       "        (relu): ReLU(inplace)\n",
       "      )\n",
       "      (2): Bottleneck(\n",
       "        (conv1): Conv2d(256, 64, kernel_size=(1, 1), stride=(1, 1), bias=False)\n",
       "        (bn1): BatchNorm2d(64, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
       "        (conv2): Conv2d(64, 64, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)\n",
       "        (bn2): BatchNorm2d(64, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
       "        (conv3): Conv2d(64, 256, kernel_size=(1, 1), stride=(1, 1), bias=False)\n",
       "        (bn3): BatchNorm2d(256, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
       "        (relu): ReLU(inplace)\n",
       "      )\n",
       "    )\n",
       "    (5): Sequential(\n",
       "      (0): Bottleneck(\n",
       "        (conv1): Conv2d(256, 128, kernel_size=(1, 1), stride=(1, 1), bias=False)\n",
       "        (bn1): BatchNorm2d(128, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
       "        (conv2): Conv2d(128, 128, kernel_size=(3, 3), stride=(2, 2), padding=(1, 1), bias=False)\n",
       "        (bn2): BatchNorm2d(128, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
       "        (conv3): Conv2d(128, 512, kernel_size=(1, 1), stride=(1, 1), bias=False)\n",
       "        (bn3): BatchNorm2d(512, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
       "        (relu): ReLU(inplace)\n",
       "        (downsample): Sequential(\n",
       "          (0): Conv2d(256, 512, kernel_size=(1, 1), stride=(2, 2), bias=False)\n",
       "          (1): BatchNorm2d(512, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
       "        )\n",
       "      )\n",
       "      (1): Bottleneck(\n",
       "        (conv1): Conv2d(512, 128, kernel_size=(1, 1), stride=(1, 1), bias=False)\n",
       "        (bn1): BatchNorm2d(128, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
       "        (conv2): Conv2d(128, 128, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)\n",
       "        (bn2): BatchNorm2d(128, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
       "        (conv3): Conv2d(128, 512, kernel_size=(1, 1), stride=(1, 1), bias=False)\n",
       "        (bn3): BatchNorm2d(512, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
       "        (relu): ReLU(inplace)\n",
       "      )\n",
       "      (2): Bottleneck(\n",
       "        (conv1): Conv2d(512, 128, kernel_size=(1, 1), stride=(1, 1), bias=False)\n",
       "        (bn1): BatchNorm2d(128, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
       "        (conv2): Conv2d(128, 128, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)\n",
       "        (bn2): BatchNorm2d(128, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
       "        (conv3): Conv2d(128, 512, kernel_size=(1, 1), stride=(1, 1), bias=False)\n",
       "        (bn3): BatchNorm2d(512, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
       "        (relu): ReLU(inplace)\n",
       "      )\n",
       "      (3): Bottleneck(\n",
       "        (conv1): Conv2d(512, 128, kernel_size=(1, 1), stride=(1, 1), bias=False)\n",
       "        (bn1): BatchNorm2d(128, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
       "        (conv2): Conv2d(128, 128, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)\n",
       "        (bn2): BatchNorm2d(128, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
       "        (conv3): Conv2d(128, 512, kernel_size=(1, 1), stride=(1, 1), bias=False)\n",
       "        (bn3): BatchNorm2d(512, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
       "        (relu): ReLU(inplace)\n",
       "      )\n",
       "    )\n",
       "    (6): Sequential(\n",
       "      (0): Bottleneck(\n",
       "        (conv1): Conv2d(512, 256, kernel_size=(1, 1), stride=(1, 1), bias=False)\n",
       "        (bn1): BatchNorm2d(256, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
       "        (conv2): Conv2d(256, 256, kernel_size=(3, 3), stride=(2, 2), padding=(1, 1), bias=False)\n",
       "        (bn2): BatchNorm2d(256, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
       "        (conv3): Conv2d(256, 1024, kernel_size=(1, 1), stride=(1, 1), bias=False)\n",
       "        (bn3): BatchNorm2d(1024, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
       "        (relu): ReLU(inplace)\n",
       "        (downsample): Sequential(\n",
       "          (0): Conv2d(512, 1024, kernel_size=(1, 1), stride=(2, 2), bias=False)\n",
       "          (1): BatchNorm2d(1024, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
       "        )\n",
       "      )\n",
       "      (1): Bottleneck(\n",
       "        (conv1): Conv2d(1024, 256, kernel_size=(1, 1), stride=(1, 1), bias=False)\n",
       "        (bn1): BatchNorm2d(256, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
       "        (conv2): Conv2d(256, 256, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)\n",
       "        (bn2): BatchNorm2d(256, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
       "        (conv3): Conv2d(256, 1024, kernel_size=(1, 1), stride=(1, 1), bias=False)\n",
       "        (bn3): BatchNorm2d(1024, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
       "        (relu): ReLU(inplace)\n",
       "      )\n",
       "      (2): Bottleneck(\n",
       "        (conv1): Conv2d(1024, 256, kernel_size=(1, 1), stride=(1, 1), bias=False)\n",
       "        (bn1): BatchNorm2d(256, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
       "        (conv2): Conv2d(256, 256, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)\n",
       "        (bn2): BatchNorm2d(256, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
       "        (conv3): Conv2d(256, 1024, kernel_size=(1, 1), stride=(1, 1), bias=False)\n",
       "        (bn3): BatchNorm2d(1024, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
       "        (relu): ReLU(inplace)\n",
       "      )\n",
       "      (3): Bottleneck(\n",
       "        (conv1): Conv2d(1024, 256, kernel_size=(1, 1), stride=(1, 1), bias=False)\n",
       "        (bn1): BatchNorm2d(256, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
       "        (conv2): Conv2d(256, 256, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)\n",
       "        (bn2): BatchNorm2d(256, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
       "        (conv3): Conv2d(256, 1024, kernel_size=(1, 1), stride=(1, 1), bias=False)\n",
       "        (bn3): BatchNorm2d(1024, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
       "        (relu): ReLU(inplace)\n",
       "      )\n",
       "      (4): Bottleneck(\n",
       "        (conv1): Conv2d(1024, 256, kernel_size=(1, 1), stride=(1, 1), bias=False)\n",
       "        (bn1): BatchNorm2d(256, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
       "        (conv2): Conv2d(256, 256, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)\n",
       "        (bn2): BatchNorm2d(256, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
       "        (conv3): Conv2d(256, 1024, kernel_size=(1, 1), stride=(1, 1), bias=False)\n",
       "        (bn3): BatchNorm2d(1024, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
       "        (relu): ReLU(inplace)\n",
       "      )\n",
       "      (5): Bottleneck(\n",
       "        (conv1): Conv2d(1024, 256, kernel_size=(1, 1), stride=(1, 1), bias=False)\n",
       "        (bn1): BatchNorm2d(256, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
       "        (conv2): Conv2d(256, 256, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)\n",
       "        (bn2): BatchNorm2d(256, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
       "        (conv3): Conv2d(256, 1024, kernel_size=(1, 1), stride=(1, 1), bias=False)\n",
       "        (bn3): BatchNorm2d(1024, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
       "        (relu): ReLU(inplace)\n",
       "      )\n",
       "      (6): Bottleneck(\n",
       "        (conv1): Conv2d(1024, 256, kernel_size=(1, 1), stride=(1, 1), bias=False)\n",
       "        (bn1): BatchNorm2d(256, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
       "        (conv2): Conv2d(256, 256, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)\n",
       "        (bn2): BatchNorm2d(256, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
       "        (conv3): Conv2d(256, 1024, kernel_size=(1, 1), stride=(1, 1), bias=False)\n",
       "        (bn3): BatchNorm2d(1024, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
       "        (relu): ReLU(inplace)\n",
       "      )\n",
       "      (7): Bottleneck(\n",
       "        (conv1): Conv2d(1024, 256, kernel_size=(1, 1), stride=(1, 1), bias=False)\n",
       "        (bn1): BatchNorm2d(256, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
       "        (conv2): Conv2d(256, 256, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)\n",
       "        (bn2): BatchNorm2d(256, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
       "        (conv3): Conv2d(256, 1024, kernel_size=(1, 1), stride=(1, 1), bias=False)\n",
       "        (bn3): BatchNorm2d(1024, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
       "        (relu): ReLU(inplace)\n",
       "      )\n",
       "      (8): Bottleneck(\n",
       "        (conv1): Conv2d(1024, 256, kernel_size=(1, 1), stride=(1, 1), bias=False)\n",
       "        (bn1): BatchNorm2d(256, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
       "        (conv2): Conv2d(256, 256, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)\n",
       "        (bn2): BatchNorm2d(256, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
       "        (conv3): Conv2d(256, 1024, kernel_size=(1, 1), stride=(1, 1), bias=False)\n",
       "        (bn3): BatchNorm2d(1024, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
       "        (relu): ReLU(inplace)\n",
       "      )\n",
       "      (9): Bottleneck(\n",
       "        (conv1): Conv2d(1024, 256, kernel_size=(1, 1), stride=(1, 1), bias=False)\n",
       "        (bn1): BatchNorm2d(256, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
       "        (conv2): Conv2d(256, 256, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)\n",
       "        (bn2): BatchNorm2d(256, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
       "        (conv3): Conv2d(256, 1024, kernel_size=(1, 1), stride=(1, 1), bias=False)\n",
       "        (bn3): BatchNorm2d(1024, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
       "        (relu): ReLU(inplace)\n",
       "      )\n",
       "      (10): Bottleneck(\n",
       "        (conv1): Conv2d(1024, 256, kernel_size=(1, 1), stride=(1, 1), bias=False)\n",
       "        (bn1): BatchNorm2d(256, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
       "        (conv2): Conv2d(256, 256, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)\n",
       "        (bn2): BatchNorm2d(256, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
       "        (conv3): Conv2d(256, 1024, kernel_size=(1, 1), stride=(1, 1), bias=False)\n",
       "        (bn3): BatchNorm2d(1024, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
       "        (relu): ReLU(inplace)\n",
       "      )\n",
       "      (11): Bottleneck(\n",
       "        (conv1): Conv2d(1024, 256, kernel_size=(1, 1), stride=(1, 1), bias=False)\n",
       "        (bn1): BatchNorm2d(256, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
       "        (conv2): Conv2d(256, 256, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)\n",
       "        (bn2): BatchNorm2d(256, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
       "        (conv3): Conv2d(256, 1024, kernel_size=(1, 1), stride=(1, 1), bias=False)\n",
       "        (bn3): BatchNorm2d(1024, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
       "        (relu): ReLU(inplace)\n",
       "      )\n",
       "      (12): Bottleneck(\n",
       "        (conv1): Conv2d(1024, 256, kernel_size=(1, 1), stride=(1, 1), bias=False)\n",
       "        (bn1): BatchNorm2d(256, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
       "        (conv2): Conv2d(256, 256, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)\n",
       "        (bn2): BatchNorm2d(256, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
       "        (conv3): Conv2d(256, 1024, kernel_size=(1, 1), stride=(1, 1), bias=False)\n",
       "        (bn3): BatchNorm2d(1024, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
       "        (relu): ReLU(inplace)\n",
       "      )\n",
       "      (13): Bottleneck(\n",
       "        (conv1): Conv2d(1024, 256, kernel_size=(1, 1), stride=(1, 1), bias=False)\n",
       "        (bn1): BatchNorm2d(256, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
       "        (conv2): Conv2d(256, 256, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)\n",
       "        (bn2): BatchNorm2d(256, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
       "        (conv3): Conv2d(256, 1024, kernel_size=(1, 1), stride=(1, 1), bias=False)\n",
       "        (bn3): BatchNorm2d(1024, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
       "        (relu): ReLU(inplace)\n",
       "      )\n",
       "      (14): Bottleneck(\n",
       "        (conv1): Conv2d(1024, 256, kernel_size=(1, 1), stride=(1, 1), bias=False)\n",
       "        (bn1): BatchNorm2d(256, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
       "        (conv2): Conv2d(256, 256, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)\n",
       "        (bn2): BatchNorm2d(256, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
       "        (conv3): Conv2d(256, 1024, kernel_size=(1, 1), stride=(1, 1), bias=False)\n",
       "        (bn3): BatchNorm2d(1024, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
       "        (relu): ReLU(inplace)\n",
       "      )\n",
       "      (15): Bottleneck(\n",
       "        (conv1): Conv2d(1024, 256, kernel_size=(1, 1), stride=(1, 1), bias=False)\n",
       "        (bn1): BatchNorm2d(256, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
       "        (conv2): Conv2d(256, 256, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)\n",
       "        (bn2): BatchNorm2d(256, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
       "        (conv3): Conv2d(256, 1024, kernel_size=(1, 1), stride=(1, 1), bias=False)\n",
       "        (bn3): BatchNorm2d(1024, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
       "        (relu): ReLU(inplace)\n",
       "      )\n",
       "      (16): Bottleneck(\n",
       "        (conv1): Conv2d(1024, 256, kernel_size=(1, 1), stride=(1, 1), bias=False)\n",
       "        (bn1): BatchNorm2d(256, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
       "        (conv2): Conv2d(256, 256, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)\n",
       "        (bn2): BatchNorm2d(256, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
       "        (conv3): Conv2d(256, 1024, kernel_size=(1, 1), stride=(1, 1), bias=False)\n",
       "        (bn3): BatchNorm2d(1024, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
       "        (relu): ReLU(inplace)\n",
       "      )\n",
       "      (17): Bottleneck(\n",
       "        (conv1): Conv2d(1024, 256, kernel_size=(1, 1), stride=(1, 1), bias=False)\n",
       "        (bn1): BatchNorm2d(256, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
       "        (conv2): Conv2d(256, 256, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)\n",
       "        (bn2): BatchNorm2d(256, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
       "        (conv3): Conv2d(256, 1024, kernel_size=(1, 1), stride=(1, 1), bias=False)\n",
       "        (bn3): BatchNorm2d(1024, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
       "        (relu): ReLU(inplace)\n",
       "      )\n",
       "      (18): Bottleneck(\n",
       "        (conv1): Conv2d(1024, 256, kernel_size=(1, 1), stride=(1, 1), bias=False)\n",
       "        (bn1): BatchNorm2d(256, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
       "        (conv2): Conv2d(256, 256, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)\n",
       "        (bn2): BatchNorm2d(256, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
       "        (conv3): Conv2d(256, 1024, kernel_size=(1, 1), stride=(1, 1), bias=False)\n",
       "        (bn3): BatchNorm2d(1024, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
       "        (relu): ReLU(inplace)\n",
       "      )\n",
       "      (19): Bottleneck(\n",
       "        (conv1): Conv2d(1024, 256, kernel_size=(1, 1), stride=(1, 1), bias=False)\n",
       "        (bn1): BatchNorm2d(256, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
       "        (conv2): Conv2d(256, 256, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)\n",
       "        (bn2): BatchNorm2d(256, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
       "        (conv3): Conv2d(256, 1024, kernel_size=(1, 1), stride=(1, 1), bias=False)\n",
       "        (bn3): BatchNorm2d(1024, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
       "        (relu): ReLU(inplace)\n",
       "      )\n",
       "      (20): Bottleneck(\n",
       "        (conv1): Conv2d(1024, 256, kernel_size=(1, 1), stride=(1, 1), bias=False)\n",
       "        (bn1): BatchNorm2d(256, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
       "        (conv2): Conv2d(256, 256, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)\n",
       "        (bn2): BatchNorm2d(256, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
       "        (conv3): Conv2d(256, 1024, kernel_size=(1, 1), stride=(1, 1), bias=False)\n",
       "        (bn3): BatchNorm2d(1024, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
       "        (relu): ReLU(inplace)\n",
       "      )\n",
       "      (21): Bottleneck(\n",
       "        (conv1): Conv2d(1024, 256, kernel_size=(1, 1), stride=(1, 1), bias=False)\n",
       "        (bn1): BatchNorm2d(256, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
       "        (conv2): Conv2d(256, 256, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)\n",
       "        (bn2): BatchNorm2d(256, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
       "        (conv3): Conv2d(256, 1024, kernel_size=(1, 1), stride=(1, 1), bias=False)\n",
       "        (bn3): BatchNorm2d(1024, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
       "        (relu): ReLU(inplace)\n",
       "      )\n",
       "      (22): Bottleneck(\n",
       "        (conv1): Conv2d(1024, 256, kernel_size=(1, 1), stride=(1, 1), bias=False)\n",
       "        (bn1): BatchNorm2d(256, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
       "        (conv2): Conv2d(256, 256, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)\n",
       "        (bn2): BatchNorm2d(256, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
       "        (conv3): Conv2d(256, 1024, kernel_size=(1, 1), stride=(1, 1), bias=False)\n",
       "        (bn3): BatchNorm2d(1024, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
       "        (relu): ReLU(inplace)\n",
       "      )\n",
       "    )\n",
       "    (7): Sequential(\n",
       "      (0): Bottleneck(\n",
       "        (conv1): Conv2d(1024, 512, kernel_size=(1, 1), stride=(1, 1), bias=False)\n",
       "        (bn1): BatchNorm2d(512, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
       "        (conv2): Conv2d(512, 512, kernel_size=(3, 3), stride=(2, 2), padding=(1, 1), bias=False)\n",
       "        (bn2): BatchNorm2d(512, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
       "        (conv3): Conv2d(512, 2048, kernel_size=(1, 1), stride=(1, 1), bias=False)\n",
       "        (bn3): BatchNorm2d(2048, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
       "        (relu): ReLU(inplace)\n",
       "        (downsample): Sequential(\n",
       "          (0): Conv2d(1024, 2048, kernel_size=(1, 1), stride=(2, 2), bias=False)\n",
       "          (1): BatchNorm2d(2048, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
       "        )\n",
       "      )\n",
       "      (1): Bottleneck(\n",
       "        (conv1): Conv2d(2048, 512, kernel_size=(1, 1), stride=(1, 1), bias=False)\n",
       "        (bn1): BatchNorm2d(512, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
       "        (conv2): Conv2d(512, 512, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)\n",
       "        (bn2): BatchNorm2d(512, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
       "        (conv3): Conv2d(512, 2048, kernel_size=(1, 1), stride=(1, 1), bias=False)\n",
       "        (bn3): BatchNorm2d(2048, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
       "        (relu): ReLU(inplace)\n",
       "      )\n",
       "      (2): Bottleneck(\n",
       "        (conv1): Conv2d(2048, 512, kernel_size=(1, 1), stride=(1, 1), bias=False)\n",
       "        (bn1): BatchNorm2d(512, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
       "        (conv2): Conv2d(512, 512, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)\n",
       "        (bn2): BatchNorm2d(512, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
       "        (conv3): Conv2d(512, 2048, kernel_size=(1, 1), stride=(1, 1), bias=False)\n",
       "        (bn3): BatchNorm2d(2048, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
       "        (relu): ReLU(inplace)\n",
       "      )\n",
       "    )\n",
       "  )\n",
       "  (1): Sequential(\n",
       "    (0): AdaptiveConcatPool2d(\n",
       "      (ap): AdaptiveAvgPool2d(output_size=1)\n",
       "      (mp): AdaptiveMaxPool2d(output_size=1)\n",
       "    )\n",
       "    (1): Flatten()\n",
       "    (2): BatchNorm1d(4096, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
       "    (3): Dropout(p=0.25)\n",
       "    (4): Linear(in_features=4096, out_features=512, bias=True)\n",
       "    (5): ReLU(inplace)\n",
       "    (6): BatchNorm1d(512, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
       "    (7): Dropout(p=0.5)\n",
       "    (8): Linear(in_features=512, out_features=1, bias=True)\n",
       "  )\n",
       "), opt_func=functools.partial(<class 'torch.optim.adam.Adam'>, betas=(0.9, 0.99)), loss_func=FlattenedLoss of MSELoss(), metrics=[<function qk at 0x7f1e1f4b9ea0>], true_wd=True, bn_wd=True, wd=0.01, train_bn=True, path=PosixPath('.'), model_dir='.', callback_fns=[functools.partial(<class 'fastai.basic_train.Recorder'>, add_time=True, silent=False), <class 'fastai.train.ShowGraph'>], callbacks=[], layer_groups=[Sequential(\n",
       "  (0): Conv2d(3, 64, kernel_size=(7, 7), stride=(2, 2), padding=(3, 3), bias=False)\n",
       "  (1): BatchNorm2d(64, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
       "  (2): ReLU(inplace)\n",
       "  (3): MaxPool2d(kernel_size=3, stride=2, padding=1, dilation=1, ceil_mode=False)\n",
       "  (4): Conv2d(64, 64, kernel_size=(1, 1), stride=(1, 1), bias=False)\n",
       "  (5): BatchNorm2d(64, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
       "  (6): Conv2d(64, 64, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)\n",
       "  (7): BatchNorm2d(64, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
       "  (8): Conv2d(64, 256, kernel_size=(1, 1), stride=(1, 1), bias=False)\n",
       "  (9): BatchNorm2d(256, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
       "  (10): ReLU(inplace)\n",
       "  (11): Conv2d(64, 256, kernel_size=(1, 1), stride=(1, 1), bias=False)\n",
       "  (12): BatchNorm2d(256, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
       "  (13): Conv2d(256, 64, kernel_size=(1, 1), stride=(1, 1), bias=False)\n",
       "  (14): BatchNorm2d(64, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
       "  (15): Conv2d(64, 64, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)\n",
       "  (16): BatchNorm2d(64, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
       "  (17): Conv2d(64, 256, kernel_size=(1, 1), stride=(1, 1), bias=False)\n",
       "  (18): BatchNorm2d(256, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
       "  (19): ReLU(inplace)\n",
       "  (20): Conv2d(256, 64, kernel_size=(1, 1), stride=(1, 1), bias=False)\n",
       "  (21): BatchNorm2d(64, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
       "  (22): Conv2d(64, 64, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)\n",
       "  (23): BatchNorm2d(64, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
       "  (24): Conv2d(64, 256, kernel_size=(1, 1), stride=(1, 1), bias=False)\n",
       "  (25): BatchNorm2d(256, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
       "  (26): ReLU(inplace)\n",
       "  (27): Conv2d(256, 128, kernel_size=(1, 1), stride=(1, 1), bias=False)\n",
       "  (28): BatchNorm2d(128, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
       "  (29): Conv2d(128, 128, kernel_size=(3, 3), stride=(2, 2), padding=(1, 1), bias=False)\n",
       "  (30): BatchNorm2d(128, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
       "  (31): Conv2d(128, 512, kernel_size=(1, 1), stride=(1, 1), bias=False)\n",
       "  (32): BatchNorm2d(512, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
       "  (33): ReLU(inplace)\n",
       "  (34): Conv2d(256, 512, kernel_size=(1, 1), stride=(2, 2), bias=False)\n",
       "  (35): BatchNorm2d(512, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
       "  (36): Conv2d(512, 128, kernel_size=(1, 1), stride=(1, 1), bias=False)\n",
       "  (37): BatchNorm2d(128, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
       "  (38): Conv2d(128, 128, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)\n",
       "  (39): BatchNorm2d(128, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
       "  (40): Conv2d(128, 512, kernel_size=(1, 1), stride=(1, 1), bias=False)\n",
       "  (41): BatchNorm2d(512, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
       "  (42): ReLU(inplace)\n",
       "  (43): Conv2d(512, 128, kernel_size=(1, 1), stride=(1, 1), bias=False)\n",
       "  (44): BatchNorm2d(128, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
       "  (45): Conv2d(128, 128, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)\n",
       "  (46): BatchNorm2d(128, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
       "  (47): Conv2d(128, 512, kernel_size=(1, 1), stride=(1, 1), bias=False)\n",
       "  (48): BatchNorm2d(512, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
       "  (49): ReLU(inplace)\n",
       "  (50): Conv2d(512, 128, kernel_size=(1, 1), stride=(1, 1), bias=False)\n",
       "  (51): BatchNorm2d(128, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
       "  (52): Conv2d(128, 128, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)\n",
       "  (53): BatchNorm2d(128, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
       "  (54): Conv2d(128, 512, kernel_size=(1, 1), stride=(1, 1), bias=False)\n",
       "  (55): BatchNorm2d(512, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
       "  (56): ReLU(inplace)\n",
       "), Sequential(\n",
       "  (0): Conv2d(512, 256, kernel_size=(1, 1), stride=(1, 1), bias=False)\n",
       "  (1): BatchNorm2d(256, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
       "  (2): Conv2d(256, 256, kernel_size=(3, 3), stride=(2, 2), padding=(1, 1), bias=False)\n",
       "  (3): BatchNorm2d(256, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
       "  (4): Conv2d(256, 1024, kernel_size=(1, 1), stride=(1, 1), bias=False)\n",
       "  (5): BatchNorm2d(1024, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
       "  (6): ReLU(inplace)\n",
       "  (7): Conv2d(512, 1024, kernel_size=(1, 1), stride=(2, 2), bias=False)\n",
       "  (8): BatchNorm2d(1024, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
       "  (9): Conv2d(1024, 256, kernel_size=(1, 1), stride=(1, 1), bias=False)\n",
       "  (10): BatchNorm2d(256, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
       "  (11): Conv2d(256, 256, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)\n",
       "  (12): BatchNorm2d(256, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
       "  (13): Conv2d(256, 1024, kernel_size=(1, 1), stride=(1, 1), bias=False)\n",
       "  (14): BatchNorm2d(1024, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
       "  (15): ReLU(inplace)\n",
       "  (16): Conv2d(1024, 256, kernel_size=(1, 1), stride=(1, 1), bias=False)\n",
       "  (17): BatchNorm2d(256, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
       "  (18): Conv2d(256, 256, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)\n",
       "  (19): BatchNorm2d(256, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
       "  (20): Conv2d(256, 1024, kernel_size=(1, 1), stride=(1, 1), bias=False)\n",
       "  (21): BatchNorm2d(1024, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
       "  (22): ReLU(inplace)\n",
       "  (23): Conv2d(1024, 256, kernel_size=(1, 1), stride=(1, 1), bias=False)\n",
       "  (24): BatchNorm2d(256, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
       "  (25): Conv2d(256, 256, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)\n",
       "  (26): BatchNorm2d(256, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
       "  (27): Conv2d(256, 1024, kernel_size=(1, 1), stride=(1, 1), bias=False)\n",
       "  (28): BatchNorm2d(1024, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
       "  (29): ReLU(inplace)\n",
       "  (30): Conv2d(1024, 256, kernel_size=(1, 1), stride=(1, 1), bias=False)\n",
       "  (31): BatchNorm2d(256, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
       "  (32): Conv2d(256, 256, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)\n",
       "  (33): BatchNorm2d(256, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
       "  (34): Conv2d(256, 1024, kernel_size=(1, 1), stride=(1, 1), bias=False)\n",
       "  (35): BatchNorm2d(1024, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
       "  (36): ReLU(inplace)\n",
       "  (37): Conv2d(1024, 256, kernel_size=(1, 1), stride=(1, 1), bias=False)\n",
       "  (38): BatchNorm2d(256, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
       "  (39): Conv2d(256, 256, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)\n",
       "  (40): BatchNorm2d(256, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
       "  (41): Conv2d(256, 1024, kernel_size=(1, 1), stride=(1, 1), bias=False)\n",
       "  (42): BatchNorm2d(1024, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
       "  (43): ReLU(inplace)\n",
       "  (44): Conv2d(1024, 256, kernel_size=(1, 1), stride=(1, 1), bias=False)\n",
       "  (45): BatchNorm2d(256, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
       "  (46): Conv2d(256, 256, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)\n",
       "  (47): BatchNorm2d(256, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
       "  (48): Conv2d(256, 1024, kernel_size=(1, 1), stride=(1, 1), bias=False)\n",
       "  (49): BatchNorm2d(1024, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
       "  (50): ReLU(inplace)\n",
       "  (51): Conv2d(1024, 256, kernel_size=(1, 1), stride=(1, 1), bias=False)\n",
       "  (52): BatchNorm2d(256, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
       "  (53): Conv2d(256, 256, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)\n",
       "  (54): BatchNorm2d(256, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
       "  (55): Conv2d(256, 1024, kernel_size=(1, 1), stride=(1, 1), bias=False)\n",
       "  (56): BatchNorm2d(1024, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
       "  (57): ReLU(inplace)\n",
       "  (58): Conv2d(1024, 256, kernel_size=(1, 1), stride=(1, 1), bias=False)\n",
       "  (59): BatchNorm2d(256, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
       "  (60): Conv2d(256, 256, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)\n",
       "  (61): BatchNorm2d(256, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
       "  (62): Conv2d(256, 1024, kernel_size=(1, 1), stride=(1, 1), bias=False)\n",
       "  (63): BatchNorm2d(1024, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
       "  (64): ReLU(inplace)\n",
       "  (65): Conv2d(1024, 256, kernel_size=(1, 1), stride=(1, 1), bias=False)\n",
       "  (66): BatchNorm2d(256, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
       "  (67): Conv2d(256, 256, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)\n",
       "  (68): BatchNorm2d(256, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
       "  (69): Conv2d(256, 1024, kernel_size=(1, 1), stride=(1, 1), bias=False)\n",
       "  (70): BatchNorm2d(1024, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
       "  (71): ReLU(inplace)\n",
       "  (72): Conv2d(1024, 256, kernel_size=(1, 1), stride=(1, 1), bias=False)\n",
       "  (73): BatchNorm2d(256, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
       "  (74): Conv2d(256, 256, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)\n",
       "  (75): BatchNorm2d(256, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
       "  (76): Conv2d(256, 1024, kernel_size=(1, 1), stride=(1, 1), bias=False)\n",
       "  (77): BatchNorm2d(1024, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
       "  (78): ReLU(inplace)\n",
       "  (79): Conv2d(1024, 256, kernel_size=(1, 1), stride=(1, 1), bias=False)\n",
       "  (80): BatchNorm2d(256, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
       "  (81): Conv2d(256, 256, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)\n",
       "  (82): BatchNorm2d(256, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
       "  (83): Conv2d(256, 1024, kernel_size=(1, 1), stride=(1, 1), bias=False)\n",
       "  (84): BatchNorm2d(1024, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
       "  (85): ReLU(inplace)\n",
       "  (86): Conv2d(1024, 256, kernel_size=(1, 1), stride=(1, 1), bias=False)\n",
       "  (87): BatchNorm2d(256, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
       "  (88): Conv2d(256, 256, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)\n",
       "  (89): BatchNorm2d(256, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
       "  (90): Conv2d(256, 1024, kernel_size=(1, 1), stride=(1, 1), bias=False)\n",
       "  (91): BatchNorm2d(1024, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
       "  (92): ReLU(inplace)\n",
       "  (93): Conv2d(1024, 256, kernel_size=(1, 1), stride=(1, 1), bias=False)\n",
       "  (94): BatchNorm2d(256, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
       "  (95): Conv2d(256, 256, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)\n",
       "  (96): BatchNorm2d(256, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
       "  (97): Conv2d(256, 1024, kernel_size=(1, 1), stride=(1, 1), bias=False)\n",
       "  (98): BatchNorm2d(1024, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
       "  (99): ReLU(inplace)\n",
       "  (100): Conv2d(1024, 256, kernel_size=(1, 1), stride=(1, 1), bias=False)\n",
       "  (101): BatchNorm2d(256, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
       "  (102): Conv2d(256, 256, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)\n",
       "  (103): BatchNorm2d(256, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
       "  (104): Conv2d(256, 1024, kernel_size=(1, 1), stride=(1, 1), bias=False)\n",
       "  (105): BatchNorm2d(1024, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
       "  (106): ReLU(inplace)\n",
       "  (107): Conv2d(1024, 256, kernel_size=(1, 1), stride=(1, 1), bias=False)\n",
       "  (108): BatchNorm2d(256, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
       "  (109): Conv2d(256, 256, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)\n",
       "  (110): BatchNorm2d(256, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
       "  (111): Conv2d(256, 1024, kernel_size=(1, 1), stride=(1, 1), bias=False)\n",
       "  (112): BatchNorm2d(1024, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
       "  (113): ReLU(inplace)\n",
       "  (114): Conv2d(1024, 256, kernel_size=(1, 1), stride=(1, 1), bias=False)\n",
       "  (115): BatchNorm2d(256, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
       "  (116): Conv2d(256, 256, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)\n",
       "  (117): BatchNorm2d(256, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
       "  (118): Conv2d(256, 1024, kernel_size=(1, 1), stride=(1, 1), bias=False)\n",
       "  (119): BatchNorm2d(1024, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
       "  (120): ReLU(inplace)\n",
       "  (121): Conv2d(1024, 256, kernel_size=(1, 1), stride=(1, 1), bias=False)\n",
       "  (122): BatchNorm2d(256, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
       "  (123): Conv2d(256, 256, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)\n",
       "  (124): BatchNorm2d(256, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
       "  (125): Conv2d(256, 1024, kernel_size=(1, 1), stride=(1, 1), bias=False)\n",
       "  (126): BatchNorm2d(1024, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
       "  (127): ReLU(inplace)\n",
       "  (128): Conv2d(1024, 256, kernel_size=(1, 1), stride=(1, 1), bias=False)\n",
       "  (129): BatchNorm2d(256, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
       "  (130): Conv2d(256, 256, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)\n",
       "  (131): BatchNorm2d(256, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
       "  (132): Conv2d(256, 1024, kernel_size=(1, 1), stride=(1, 1), bias=False)\n",
       "  (133): BatchNorm2d(1024, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
       "  (134): ReLU(inplace)\n",
       "  (135): Conv2d(1024, 256, kernel_size=(1, 1), stride=(1, 1), bias=False)\n",
       "  (136): BatchNorm2d(256, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
       "  (137): Conv2d(256, 256, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)\n",
       "  (138): BatchNorm2d(256, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
       "  (139): Conv2d(256, 1024, kernel_size=(1, 1), stride=(1, 1), bias=False)\n",
       "  (140): BatchNorm2d(1024, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
       "  (141): ReLU(inplace)\n",
       "  (142): Conv2d(1024, 256, kernel_size=(1, 1), stride=(1, 1), bias=False)\n",
       "  (143): BatchNorm2d(256, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
       "  (144): Conv2d(256, 256, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)\n",
       "  (145): BatchNorm2d(256, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
       "  (146): Conv2d(256, 1024, kernel_size=(1, 1), stride=(1, 1), bias=False)\n",
       "  (147): BatchNorm2d(1024, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
       "  (148): ReLU(inplace)\n",
       "  (149): Conv2d(1024, 256, kernel_size=(1, 1), stride=(1, 1), bias=False)\n",
       "  (150): BatchNorm2d(256, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
       "  (151): Conv2d(256, 256, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)\n",
       "  (152): BatchNorm2d(256, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
       "  (153): Conv2d(256, 1024, kernel_size=(1, 1), stride=(1, 1), bias=False)\n",
       "  (154): BatchNorm2d(1024, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
       "  (155): ReLU(inplace)\n",
       "  (156): Conv2d(1024, 256, kernel_size=(1, 1), stride=(1, 1), bias=False)\n",
       "  (157): BatchNorm2d(256, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
       "  (158): Conv2d(256, 256, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)\n",
       "  (159): BatchNorm2d(256, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
       "  (160): Conv2d(256, 1024, kernel_size=(1, 1), stride=(1, 1), bias=False)\n",
       "  (161): BatchNorm2d(1024, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
       "  (162): ReLU(inplace)\n",
       "  (163): Conv2d(1024, 512, kernel_size=(1, 1), stride=(1, 1), bias=False)\n",
       "  (164): BatchNorm2d(512, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
       "  (165): Conv2d(512, 512, kernel_size=(3, 3), stride=(2, 2), padding=(1, 1), bias=False)\n",
       "  (166): BatchNorm2d(512, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
       "  (167): Conv2d(512, 2048, kernel_size=(1, 1), stride=(1, 1), bias=False)\n",
       "  (168): BatchNorm2d(2048, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
       "  (169): ReLU(inplace)\n",
       "  (170): Conv2d(1024, 2048, kernel_size=(1, 1), stride=(2, 2), bias=False)\n",
       "  (171): BatchNorm2d(2048, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
       "  (172): Conv2d(2048, 512, kernel_size=(1, 1), stride=(1, 1), bias=False)\n",
       "  (173): BatchNorm2d(512, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
       "  (174): Conv2d(512, 512, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)\n",
       "  (175): BatchNorm2d(512, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
       "  (176): Conv2d(512, 2048, kernel_size=(1, 1), stride=(1, 1), bias=False)\n",
       "  (177): BatchNorm2d(2048, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
       "  (178): ReLU(inplace)\n",
       "  (179): Conv2d(2048, 512, kernel_size=(1, 1), stride=(1, 1), bias=False)\n",
       "  (180): BatchNorm2d(512, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
       "  (181): Conv2d(512, 512, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)\n",
       "  (182): BatchNorm2d(512, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
       "  (183): Conv2d(512, 2048, kernel_size=(1, 1), stride=(1, 1), bias=False)\n",
       "  (184): BatchNorm2d(2048, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
       "  (185): ReLU(inplace)\n",
       "), Sequential(\n",
       "  (0): AdaptiveAvgPool2d(output_size=1)\n",
       "  (1): AdaptiveMaxPool2d(output_size=1)\n",
       "  (2): Flatten()\n",
       "  (3): BatchNorm1d(4096, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
       "  (4): Dropout(p=0.25)\n",
       "  (5): Linear(in_features=4096, out_features=512, bias=True)\n",
       "  (6): ReLU(inplace)\n",
       "  (7): BatchNorm1d(512, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
       "  (8): Dropout(p=0.5)\n",
       "  (9): Linear(in_features=512, out_features=1, bias=True)\n",
       ")], add_time=True, silent=False)"
      ]
     },
     "execution_count": 17,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "learn.load('best_model')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 18,
   "metadata": {},
   "outputs": [],
   "source": [
    "learn.unfreeze()\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 19,
   "metadata": {},
   "outputs": [],
   "source": [
    "# learn.lr_find()\n",
    "# learn.recorder.plot()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 20,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: left;\">\n",
       "      <th>epoch</th>\n",
       "      <th>train_loss</th>\n",
       "      <th>valid_loss</th>\n",
       "      <th>qk</th>\n",
       "      <th>time</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <td>0</td>\n",
       "      <td>0.330040</td>\n",
       "      <td>0.284293</td>\n",
       "      <td>0.889500</td>\n",
       "      <td>04:43</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>1</td>\n",
       "      <td>0.336030</td>\n",
       "      <td>0.283969</td>\n",
       "      <td>0.902223</td>\n",
       "      <td>04:27</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>2</td>\n",
       "      <td>0.336175</td>\n",
       "      <td>0.296801</td>\n",
       "      <td>0.893939</td>\n",
       "      <td>04:32</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>3</td>\n",
       "      <td>0.339237</td>\n",
       "      <td>0.284736</td>\n",
       "      <td>0.887854</td>\n",
       "      <td>04:47</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>4</td>\n",
       "      <td>0.316380</td>\n",
       "      <td>0.284324</td>\n",
       "      <td>0.893015</td>\n",
       "      <td>04:48</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>5</td>\n",
       "      <td>0.312376</td>\n",
       "      <td>0.282387</td>\n",
       "      <td>0.900477</td>\n",
       "      <td>04:44</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>6</td>\n",
       "      <td>0.312719</td>\n",
       "      <td>0.284831</td>\n",
       "      <td>0.890136</td>\n",
       "      <td>04:47</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>7</td>\n",
       "      <td>0.319583</td>\n",
       "      <td>0.290804</td>\n",
       "      <td>0.890237</td>\n",
       "      <td>04:43</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>8</td>\n",
       "      <td>0.294826</td>\n",
       "      <td>0.282518</td>\n",
       "      <td>0.894105</td>\n",
       "      <td>04:28</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>9</td>\n",
       "      <td>0.326056</td>\n",
       "      <td>0.279302</td>\n",
       "      <td>0.897693</td>\n",
       "      <td>04:27</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>"
      ],
      "text/plain": [
       "<IPython.core.display.HTML object>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Better model found at epoch 0 with qk value: 0.8895004987716675.\n",
      "Better model found at epoch 1 with qk value: 0.9022226929664612.\n"
     ]
    }
   ],
   "source": [
    "learn.fit_one_cycle(10, max_lr=slice(1e-6,1e-4), callbacks=[callbacks.SaveModelCallback(learn,monitor='qk',mode='max', name='best_model')])\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 21,
   "metadata": {},
   "outputs": [],
   "source": [
    "#learn.load('best_model')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 22,
   "metadata": {},
   "outputs": [],
   "source": [
    "#learn.save('resnet_50_stage-1')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 23,
   "metadata": {},
   "outputs": [],
   "source": [
    "## Progressive Resizing of the Image"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 24,
   "metadata": {},
   "outputs": [],
   "source": [
    "# data = (\n",
    "#     train.split_by_rand_pct(0.2)\n",
    "#     .label_from_df(cols='diagnosis', label_cls=FloatList)\n",
    "#     .add_test(test)\n",
    "#     .transform(tfms, size=256)\n",
    "#     .databunch(path=Path('.'), bs=32).normalize(imagenet_stats)\n",
    "# )"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 25,
   "metadata": {},
   "outputs": [],
   "source": [
    "# learn = cnn_learner(data, models.resnet50, metrics=[qk], model_dir=\".\", callback_fns=ShowGraph)\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 26,
   "metadata": {},
   "outputs": [],
   "source": [
    "#learn.load('resnet_50_stage-1')\n",
    "# learn.load('best_model')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 27,
   "metadata": {},
   "outputs": [],
   "source": [
    "#learn.lr_find()\n",
    "#learn.recorder.plot()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 28,
   "metadata": {},
   "outputs": [],
   "source": [
    "# learn.fit_one_cycle(10, 1e-3, callbacks=[callbacks.SaveModelCallback(learn,monitor='qk',mode='max', name='best_model')])\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 29,
   "metadata": {},
   "outputs": [],
   "source": [
    "# learn.load('best_model')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 30,
   "metadata": {},
   "outputs": [],
   "source": [
    "# learn.unfreeze()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 31,
   "metadata": {},
   "outputs": [],
   "source": [
    "# learn.lr_find()\n",
    "# learn.recorder.plot()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 32,
   "metadata": {},
   "outputs": [],
   "source": [
    "# learn.fit_one_cycle(10, max_lr=slice(3e-6,2e-3), callbacks=[callbacks.SaveModelCallback(learn,monitor='qk',mode='max')])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 33,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "Learner(data=ImageDataBunch;\n",
       "\n",
       "Train: LabelList (2930 items)\n",
       "x: ImageList\n",
       "Image (1, 224, 224),Image (1, 224, 224),Image (1, 224, 224),Image (1, 224, 224),Image (1, 224, 224)\n",
       "y: FloatList\n",
       "2.0,4.0,1.0,0.0,0.0\n",
       "Path: ../input/aptos2019-blindness-detection;\n",
       "\n",
       "Valid: LabelList (732 items)\n",
       "x: ImageList\n",
       "Image (1, 224, 224),Image (1, 224, 224),Image (1, 224, 224),Image (1, 224, 224),Image (1, 224, 224)\n",
       "y: FloatList\n",
       "0.0,2.0,0.0,0.0,0.0\n",
       "Path: ../input/aptos2019-blindness-detection;\n",
       "\n",
       "Test: LabelList (1928 items)\n",
       "x: ImageList\n",
       "Image (1, 224, 224),Image (1, 224, 224),Image (1, 224, 224),Image (1, 224, 224),Image (1, 224, 224)\n",
       "y: EmptyLabelList\n",
       ",,,,\n",
       "Path: ../input/aptos2019-blindness-detection, model=Sequential(\n",
       "  (0): Sequential(\n",
       "    (0): Conv2d(3, 64, kernel_size=(7, 7), stride=(2, 2), padding=(3, 3), bias=False)\n",
       "    (1): BatchNorm2d(64, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
       "    (2): ReLU(inplace)\n",
       "    (3): MaxPool2d(kernel_size=3, stride=2, padding=1, dilation=1, ceil_mode=False)\n",
       "    (4): Sequential(\n",
       "      (0): Bottleneck(\n",
       "        (conv1): Conv2d(64, 64, kernel_size=(1, 1), stride=(1, 1), bias=False)\n",
       "        (bn1): BatchNorm2d(64, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
       "        (conv2): Conv2d(64, 64, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)\n",
       "        (bn2): BatchNorm2d(64, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
       "        (conv3): Conv2d(64, 256, kernel_size=(1, 1), stride=(1, 1), bias=False)\n",
       "        (bn3): BatchNorm2d(256, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
       "        (relu): ReLU(inplace)\n",
       "        (downsample): Sequential(\n",
       "          (0): Conv2d(64, 256, kernel_size=(1, 1), stride=(1, 1), bias=False)\n",
       "          (1): BatchNorm2d(256, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
       "        )\n",
       "      )\n",
       "      (1): Bottleneck(\n",
       "        (conv1): Conv2d(256, 64, kernel_size=(1, 1), stride=(1, 1), bias=False)\n",
       "        (bn1): BatchNorm2d(64, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
       "        (conv2): Conv2d(64, 64, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)\n",
       "        (bn2): BatchNorm2d(64, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
       "        (conv3): Conv2d(64, 256, kernel_size=(1, 1), stride=(1, 1), bias=False)\n",
       "        (bn3): BatchNorm2d(256, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
       "        (relu): ReLU(inplace)\n",
       "      )\n",
       "      (2): Bottleneck(\n",
       "        (conv1): Conv2d(256, 64, kernel_size=(1, 1), stride=(1, 1), bias=False)\n",
       "        (bn1): BatchNorm2d(64, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
       "        (conv2): Conv2d(64, 64, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)\n",
       "        (bn2): BatchNorm2d(64, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
       "        (conv3): Conv2d(64, 256, kernel_size=(1, 1), stride=(1, 1), bias=False)\n",
       "        (bn3): BatchNorm2d(256, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
       "        (relu): ReLU(inplace)\n",
       "      )\n",
       "    )\n",
       "    (5): Sequential(\n",
       "      (0): Bottleneck(\n",
       "        (conv1): Conv2d(256, 128, kernel_size=(1, 1), stride=(1, 1), bias=False)\n",
       "        (bn1): BatchNorm2d(128, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
       "        (conv2): Conv2d(128, 128, kernel_size=(3, 3), stride=(2, 2), padding=(1, 1), bias=False)\n",
       "        (bn2): BatchNorm2d(128, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
       "        (conv3): Conv2d(128, 512, kernel_size=(1, 1), stride=(1, 1), bias=False)\n",
       "        (bn3): BatchNorm2d(512, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
       "        (relu): ReLU(inplace)\n",
       "        (downsample): Sequential(\n",
       "          (0): Conv2d(256, 512, kernel_size=(1, 1), stride=(2, 2), bias=False)\n",
       "          (1): BatchNorm2d(512, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
       "        )\n",
       "      )\n",
       "      (1): Bottleneck(\n",
       "        (conv1): Conv2d(512, 128, kernel_size=(1, 1), stride=(1, 1), bias=False)\n",
       "        (bn1): BatchNorm2d(128, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
       "        (conv2): Conv2d(128, 128, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)\n",
       "        (bn2): BatchNorm2d(128, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
       "        (conv3): Conv2d(128, 512, kernel_size=(1, 1), stride=(1, 1), bias=False)\n",
       "        (bn3): BatchNorm2d(512, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
       "        (relu): ReLU(inplace)\n",
       "      )\n",
       "      (2): Bottleneck(\n",
       "        (conv1): Conv2d(512, 128, kernel_size=(1, 1), stride=(1, 1), bias=False)\n",
       "        (bn1): BatchNorm2d(128, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
       "        (conv2): Conv2d(128, 128, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)\n",
       "        (bn2): BatchNorm2d(128, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
       "        (conv3): Conv2d(128, 512, kernel_size=(1, 1), stride=(1, 1), bias=False)\n",
       "        (bn3): BatchNorm2d(512, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
       "        (relu): ReLU(inplace)\n",
       "      )\n",
       "      (3): Bottleneck(\n",
       "        (conv1): Conv2d(512, 128, kernel_size=(1, 1), stride=(1, 1), bias=False)\n",
       "        (bn1): BatchNorm2d(128, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
       "        (conv2): Conv2d(128, 128, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)\n",
       "        (bn2): BatchNorm2d(128, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
       "        (conv3): Conv2d(128, 512, kernel_size=(1, 1), stride=(1, 1), bias=False)\n",
       "        (bn3): BatchNorm2d(512, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
       "        (relu): ReLU(inplace)\n",
       "      )\n",
       "    )\n",
       "    (6): Sequential(\n",
       "      (0): Bottleneck(\n",
       "        (conv1): Conv2d(512, 256, kernel_size=(1, 1), stride=(1, 1), bias=False)\n",
       "        (bn1): BatchNorm2d(256, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
       "        (conv2): Conv2d(256, 256, kernel_size=(3, 3), stride=(2, 2), padding=(1, 1), bias=False)\n",
       "        (bn2): BatchNorm2d(256, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
       "        (conv3): Conv2d(256, 1024, kernel_size=(1, 1), stride=(1, 1), bias=False)\n",
       "        (bn3): BatchNorm2d(1024, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
       "        (relu): ReLU(inplace)\n",
       "        (downsample): Sequential(\n",
       "          (0): Conv2d(512, 1024, kernel_size=(1, 1), stride=(2, 2), bias=False)\n",
       "          (1): BatchNorm2d(1024, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
       "        )\n",
       "      )\n",
       "      (1): Bottleneck(\n",
       "        (conv1): Conv2d(1024, 256, kernel_size=(1, 1), stride=(1, 1), bias=False)\n",
       "        (bn1): BatchNorm2d(256, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
       "        (conv2): Conv2d(256, 256, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)\n",
       "        (bn2): BatchNorm2d(256, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
       "        (conv3): Conv2d(256, 1024, kernel_size=(1, 1), stride=(1, 1), bias=False)\n",
       "        (bn3): BatchNorm2d(1024, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
       "        (relu): ReLU(inplace)\n",
       "      )\n",
       "      (2): Bottleneck(\n",
       "        (conv1): Conv2d(1024, 256, kernel_size=(1, 1), stride=(1, 1), bias=False)\n",
       "        (bn1): BatchNorm2d(256, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
       "        (conv2): Conv2d(256, 256, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)\n",
       "        (bn2): BatchNorm2d(256, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
       "        (conv3): Conv2d(256, 1024, kernel_size=(1, 1), stride=(1, 1), bias=False)\n",
       "        (bn3): BatchNorm2d(1024, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
       "        (relu): ReLU(inplace)\n",
       "      )\n",
       "      (3): Bottleneck(\n",
       "        (conv1): Conv2d(1024, 256, kernel_size=(1, 1), stride=(1, 1), bias=False)\n",
       "        (bn1): BatchNorm2d(256, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
       "        (conv2): Conv2d(256, 256, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)\n",
       "        (bn2): BatchNorm2d(256, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
       "        (conv3): Conv2d(256, 1024, kernel_size=(1, 1), stride=(1, 1), bias=False)\n",
       "        (bn3): BatchNorm2d(1024, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
       "        (relu): ReLU(inplace)\n",
       "      )\n",
       "      (4): Bottleneck(\n",
       "        (conv1): Conv2d(1024, 256, kernel_size=(1, 1), stride=(1, 1), bias=False)\n",
       "        (bn1): BatchNorm2d(256, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
       "        (conv2): Conv2d(256, 256, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)\n",
       "        (bn2): BatchNorm2d(256, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
       "        (conv3): Conv2d(256, 1024, kernel_size=(1, 1), stride=(1, 1), bias=False)\n",
       "        (bn3): BatchNorm2d(1024, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
       "        (relu): ReLU(inplace)\n",
       "      )\n",
       "      (5): Bottleneck(\n",
       "        (conv1): Conv2d(1024, 256, kernel_size=(1, 1), stride=(1, 1), bias=False)\n",
       "        (bn1): BatchNorm2d(256, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
       "        (conv2): Conv2d(256, 256, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)\n",
       "        (bn2): BatchNorm2d(256, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
       "        (conv3): Conv2d(256, 1024, kernel_size=(1, 1), stride=(1, 1), bias=False)\n",
       "        (bn3): BatchNorm2d(1024, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
       "        (relu): ReLU(inplace)\n",
       "      )\n",
       "      (6): Bottleneck(\n",
       "        (conv1): Conv2d(1024, 256, kernel_size=(1, 1), stride=(1, 1), bias=False)\n",
       "        (bn1): BatchNorm2d(256, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
       "        (conv2): Conv2d(256, 256, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)\n",
       "        (bn2): BatchNorm2d(256, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
       "        (conv3): Conv2d(256, 1024, kernel_size=(1, 1), stride=(1, 1), bias=False)\n",
       "        (bn3): BatchNorm2d(1024, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
       "        (relu): ReLU(inplace)\n",
       "      )\n",
       "      (7): Bottleneck(\n",
       "        (conv1): Conv2d(1024, 256, kernel_size=(1, 1), stride=(1, 1), bias=False)\n",
       "        (bn1): BatchNorm2d(256, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
       "        (conv2): Conv2d(256, 256, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)\n",
       "        (bn2): BatchNorm2d(256, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
       "        (conv3): Conv2d(256, 1024, kernel_size=(1, 1), stride=(1, 1), bias=False)\n",
       "        (bn3): BatchNorm2d(1024, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
       "        (relu): ReLU(inplace)\n",
       "      )\n",
       "      (8): Bottleneck(\n",
       "        (conv1): Conv2d(1024, 256, kernel_size=(1, 1), stride=(1, 1), bias=False)\n",
       "        (bn1): BatchNorm2d(256, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
       "        (conv2): Conv2d(256, 256, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)\n",
       "        (bn2): BatchNorm2d(256, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
       "        (conv3): Conv2d(256, 1024, kernel_size=(1, 1), stride=(1, 1), bias=False)\n",
       "        (bn3): BatchNorm2d(1024, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
       "        (relu): ReLU(inplace)\n",
       "      )\n",
       "      (9): Bottleneck(\n",
       "        (conv1): Conv2d(1024, 256, kernel_size=(1, 1), stride=(1, 1), bias=False)\n",
       "        (bn1): BatchNorm2d(256, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
       "        (conv2): Conv2d(256, 256, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)\n",
       "        (bn2): BatchNorm2d(256, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
       "        (conv3): Conv2d(256, 1024, kernel_size=(1, 1), stride=(1, 1), bias=False)\n",
       "        (bn3): BatchNorm2d(1024, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
       "        (relu): ReLU(inplace)\n",
       "      )\n",
       "      (10): Bottleneck(\n",
       "        (conv1): Conv2d(1024, 256, kernel_size=(1, 1), stride=(1, 1), bias=False)\n",
       "        (bn1): BatchNorm2d(256, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
       "        (conv2): Conv2d(256, 256, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)\n",
       "        (bn2): BatchNorm2d(256, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
       "        (conv3): Conv2d(256, 1024, kernel_size=(1, 1), stride=(1, 1), bias=False)\n",
       "        (bn3): BatchNorm2d(1024, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
       "        (relu): ReLU(inplace)\n",
       "      )\n",
       "      (11): Bottleneck(\n",
       "        (conv1): Conv2d(1024, 256, kernel_size=(1, 1), stride=(1, 1), bias=False)\n",
       "        (bn1): BatchNorm2d(256, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
       "        (conv2): Conv2d(256, 256, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)\n",
       "        (bn2): BatchNorm2d(256, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
       "        (conv3): Conv2d(256, 1024, kernel_size=(1, 1), stride=(1, 1), bias=False)\n",
       "        (bn3): BatchNorm2d(1024, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
       "        (relu): ReLU(inplace)\n",
       "      )\n",
       "      (12): Bottleneck(\n",
       "        (conv1): Conv2d(1024, 256, kernel_size=(1, 1), stride=(1, 1), bias=False)\n",
       "        (bn1): BatchNorm2d(256, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
       "        (conv2): Conv2d(256, 256, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)\n",
       "        (bn2): BatchNorm2d(256, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
       "        (conv3): Conv2d(256, 1024, kernel_size=(1, 1), stride=(1, 1), bias=False)\n",
       "        (bn3): BatchNorm2d(1024, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
       "        (relu): ReLU(inplace)\n",
       "      )\n",
       "      (13): Bottleneck(\n",
       "        (conv1): Conv2d(1024, 256, kernel_size=(1, 1), stride=(1, 1), bias=False)\n",
       "        (bn1): BatchNorm2d(256, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
       "        (conv2): Conv2d(256, 256, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)\n",
       "        (bn2): BatchNorm2d(256, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
       "        (conv3): Conv2d(256, 1024, kernel_size=(1, 1), stride=(1, 1), bias=False)\n",
       "        (bn3): BatchNorm2d(1024, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
       "        (relu): ReLU(inplace)\n",
       "      )\n",
       "      (14): Bottleneck(\n",
       "        (conv1): Conv2d(1024, 256, kernel_size=(1, 1), stride=(1, 1), bias=False)\n",
       "        (bn1): BatchNorm2d(256, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
       "        (conv2): Conv2d(256, 256, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)\n",
       "        (bn2): BatchNorm2d(256, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
       "        (conv3): Conv2d(256, 1024, kernel_size=(1, 1), stride=(1, 1), bias=False)\n",
       "        (bn3): BatchNorm2d(1024, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
       "        (relu): ReLU(inplace)\n",
       "      )\n",
       "      (15): Bottleneck(\n",
       "        (conv1): Conv2d(1024, 256, kernel_size=(1, 1), stride=(1, 1), bias=False)\n",
       "        (bn1): BatchNorm2d(256, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
       "        (conv2): Conv2d(256, 256, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)\n",
       "        (bn2): BatchNorm2d(256, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
       "        (conv3): Conv2d(256, 1024, kernel_size=(1, 1), stride=(1, 1), bias=False)\n",
       "        (bn3): BatchNorm2d(1024, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
       "        (relu): ReLU(inplace)\n",
       "      )\n",
       "      (16): Bottleneck(\n",
       "        (conv1): Conv2d(1024, 256, kernel_size=(1, 1), stride=(1, 1), bias=False)\n",
       "        (bn1): BatchNorm2d(256, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
       "        (conv2): Conv2d(256, 256, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)\n",
       "        (bn2): BatchNorm2d(256, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
       "        (conv3): Conv2d(256, 1024, kernel_size=(1, 1), stride=(1, 1), bias=False)\n",
       "        (bn3): BatchNorm2d(1024, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
       "        (relu): ReLU(inplace)\n",
       "      )\n",
       "      (17): Bottleneck(\n",
       "        (conv1): Conv2d(1024, 256, kernel_size=(1, 1), stride=(1, 1), bias=False)\n",
       "        (bn1): BatchNorm2d(256, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
       "        (conv2): Conv2d(256, 256, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)\n",
       "        (bn2): BatchNorm2d(256, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
       "        (conv3): Conv2d(256, 1024, kernel_size=(1, 1), stride=(1, 1), bias=False)\n",
       "        (bn3): BatchNorm2d(1024, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
       "        (relu): ReLU(inplace)\n",
       "      )\n",
       "      (18): Bottleneck(\n",
       "        (conv1): Conv2d(1024, 256, kernel_size=(1, 1), stride=(1, 1), bias=False)\n",
       "        (bn1): BatchNorm2d(256, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
       "        (conv2): Conv2d(256, 256, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)\n",
       "        (bn2): BatchNorm2d(256, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
       "        (conv3): Conv2d(256, 1024, kernel_size=(1, 1), stride=(1, 1), bias=False)\n",
       "        (bn3): BatchNorm2d(1024, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
       "        (relu): ReLU(inplace)\n",
       "      )\n",
       "      (19): Bottleneck(\n",
       "        (conv1): Conv2d(1024, 256, kernel_size=(1, 1), stride=(1, 1), bias=False)\n",
       "        (bn1): BatchNorm2d(256, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
       "        (conv2): Conv2d(256, 256, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)\n",
       "        (bn2): BatchNorm2d(256, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
       "        (conv3): Conv2d(256, 1024, kernel_size=(1, 1), stride=(1, 1), bias=False)\n",
       "        (bn3): BatchNorm2d(1024, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
       "        (relu): ReLU(inplace)\n",
       "      )\n",
       "      (20): Bottleneck(\n",
       "        (conv1): Conv2d(1024, 256, kernel_size=(1, 1), stride=(1, 1), bias=False)\n",
       "        (bn1): BatchNorm2d(256, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
       "        (conv2): Conv2d(256, 256, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)\n",
       "        (bn2): BatchNorm2d(256, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
       "        (conv3): Conv2d(256, 1024, kernel_size=(1, 1), stride=(1, 1), bias=False)\n",
       "        (bn3): BatchNorm2d(1024, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
       "        (relu): ReLU(inplace)\n",
       "      )\n",
       "      (21): Bottleneck(\n",
       "        (conv1): Conv2d(1024, 256, kernel_size=(1, 1), stride=(1, 1), bias=False)\n",
       "        (bn1): BatchNorm2d(256, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
       "        (conv2): Conv2d(256, 256, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)\n",
       "        (bn2): BatchNorm2d(256, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
       "        (conv3): Conv2d(256, 1024, kernel_size=(1, 1), stride=(1, 1), bias=False)\n",
       "        (bn3): BatchNorm2d(1024, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
       "        (relu): ReLU(inplace)\n",
       "      )\n",
       "      (22): Bottleneck(\n",
       "        (conv1): Conv2d(1024, 256, kernel_size=(1, 1), stride=(1, 1), bias=False)\n",
       "        (bn1): BatchNorm2d(256, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
       "        (conv2): Conv2d(256, 256, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)\n",
       "        (bn2): BatchNorm2d(256, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
       "        (conv3): Conv2d(256, 1024, kernel_size=(1, 1), stride=(1, 1), bias=False)\n",
       "        (bn3): BatchNorm2d(1024, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
       "        (relu): ReLU(inplace)\n",
       "      )\n",
       "    )\n",
       "    (7): Sequential(\n",
       "      (0): Bottleneck(\n",
       "        (conv1): Conv2d(1024, 512, kernel_size=(1, 1), stride=(1, 1), bias=False)\n",
       "        (bn1): BatchNorm2d(512, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
       "        (conv2): Conv2d(512, 512, kernel_size=(3, 3), stride=(2, 2), padding=(1, 1), bias=False)\n",
       "        (bn2): BatchNorm2d(512, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
       "        (conv3): Conv2d(512, 2048, kernel_size=(1, 1), stride=(1, 1), bias=False)\n",
       "        (bn3): BatchNorm2d(2048, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
       "        (relu): ReLU(inplace)\n",
       "        (downsample): Sequential(\n",
       "          (0): Conv2d(1024, 2048, kernel_size=(1, 1), stride=(2, 2), bias=False)\n",
       "          (1): BatchNorm2d(2048, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
       "        )\n",
       "      )\n",
       "      (1): Bottleneck(\n",
       "        (conv1): Conv2d(2048, 512, kernel_size=(1, 1), stride=(1, 1), bias=False)\n",
       "        (bn1): BatchNorm2d(512, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
       "        (conv2): Conv2d(512, 512, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)\n",
       "        (bn2): BatchNorm2d(512, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
       "        (conv3): Conv2d(512, 2048, kernel_size=(1, 1), stride=(1, 1), bias=False)\n",
       "        (bn3): BatchNorm2d(2048, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
       "        (relu): ReLU(inplace)\n",
       "      )\n",
       "      (2): Bottleneck(\n",
       "        (conv1): Conv2d(2048, 512, kernel_size=(1, 1), stride=(1, 1), bias=False)\n",
       "        (bn1): BatchNorm2d(512, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
       "        (conv2): Conv2d(512, 512, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)\n",
       "        (bn2): BatchNorm2d(512, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
       "        (conv3): Conv2d(512, 2048, kernel_size=(1, 1), stride=(1, 1), bias=False)\n",
       "        (bn3): BatchNorm2d(2048, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
       "        (relu): ReLU(inplace)\n",
       "      )\n",
       "    )\n",
       "  )\n",
       "  (1): Sequential(\n",
       "    (0): AdaptiveConcatPool2d(\n",
       "      (ap): AdaptiveAvgPool2d(output_size=1)\n",
       "      (mp): AdaptiveMaxPool2d(output_size=1)\n",
       "    )\n",
       "    (1): Flatten()\n",
       "    (2): BatchNorm1d(4096, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
       "    (3): Dropout(p=0.25)\n",
       "    (4): Linear(in_features=4096, out_features=512, bias=True)\n",
       "    (5): ReLU(inplace)\n",
       "    (6): BatchNorm1d(512, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
       "    (7): Dropout(p=0.5)\n",
       "    (8): Linear(in_features=512, out_features=1, bias=True)\n",
       "  )\n",
       "), opt_func=functools.partial(<class 'torch.optim.adam.Adam'>, betas=(0.9, 0.99)), loss_func=FlattenedLoss of MSELoss(), metrics=[<function qk at 0x7f1e1f4b9ea0>], true_wd=True, bn_wd=True, wd=0.01, train_bn=True, path=PosixPath('.'), model_dir='.', callback_fns=[functools.partial(<class 'fastai.basic_train.Recorder'>, add_time=True, silent=False), <class 'fastai.train.ShowGraph'>], callbacks=[], layer_groups=[Sequential(\n",
       "  (0): Conv2d(3, 64, kernel_size=(7, 7), stride=(2, 2), padding=(3, 3), bias=False)\n",
       "  (1): BatchNorm2d(64, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
       "  (2): ReLU(inplace)\n",
       "  (3): MaxPool2d(kernel_size=3, stride=2, padding=1, dilation=1, ceil_mode=False)\n",
       "  (4): Conv2d(64, 64, kernel_size=(1, 1), stride=(1, 1), bias=False)\n",
       "  (5): BatchNorm2d(64, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
       "  (6): Conv2d(64, 64, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)\n",
       "  (7): BatchNorm2d(64, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
       "  (8): Conv2d(64, 256, kernel_size=(1, 1), stride=(1, 1), bias=False)\n",
       "  (9): BatchNorm2d(256, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
       "  (10): ReLU(inplace)\n",
       "  (11): Conv2d(64, 256, kernel_size=(1, 1), stride=(1, 1), bias=False)\n",
       "  (12): BatchNorm2d(256, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
       "  (13): Conv2d(256, 64, kernel_size=(1, 1), stride=(1, 1), bias=False)\n",
       "  (14): BatchNorm2d(64, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
       "  (15): Conv2d(64, 64, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)\n",
       "  (16): BatchNorm2d(64, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
       "  (17): Conv2d(64, 256, kernel_size=(1, 1), stride=(1, 1), bias=False)\n",
       "  (18): BatchNorm2d(256, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
       "  (19): ReLU(inplace)\n",
       "  (20): Conv2d(256, 64, kernel_size=(1, 1), stride=(1, 1), bias=False)\n",
       "  (21): BatchNorm2d(64, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
       "  (22): Conv2d(64, 64, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)\n",
       "  (23): BatchNorm2d(64, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
       "  (24): Conv2d(64, 256, kernel_size=(1, 1), stride=(1, 1), bias=False)\n",
       "  (25): BatchNorm2d(256, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
       "  (26): ReLU(inplace)\n",
       "  (27): Conv2d(256, 128, kernel_size=(1, 1), stride=(1, 1), bias=False)\n",
       "  (28): BatchNorm2d(128, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
       "  (29): Conv2d(128, 128, kernel_size=(3, 3), stride=(2, 2), padding=(1, 1), bias=False)\n",
       "  (30): BatchNorm2d(128, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
       "  (31): Conv2d(128, 512, kernel_size=(1, 1), stride=(1, 1), bias=False)\n",
       "  (32): BatchNorm2d(512, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
       "  (33): ReLU(inplace)\n",
       "  (34): Conv2d(256, 512, kernel_size=(1, 1), stride=(2, 2), bias=False)\n",
       "  (35): BatchNorm2d(512, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
       "  (36): Conv2d(512, 128, kernel_size=(1, 1), stride=(1, 1), bias=False)\n",
       "  (37): BatchNorm2d(128, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
       "  (38): Conv2d(128, 128, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)\n",
       "  (39): BatchNorm2d(128, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
       "  (40): Conv2d(128, 512, kernel_size=(1, 1), stride=(1, 1), bias=False)\n",
       "  (41): BatchNorm2d(512, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
       "  (42): ReLU(inplace)\n",
       "  (43): Conv2d(512, 128, kernel_size=(1, 1), stride=(1, 1), bias=False)\n",
       "  (44): BatchNorm2d(128, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
       "  (45): Conv2d(128, 128, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)\n",
       "  (46): BatchNorm2d(128, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
       "  (47): Conv2d(128, 512, kernel_size=(1, 1), stride=(1, 1), bias=False)\n",
       "  (48): BatchNorm2d(512, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
       "  (49): ReLU(inplace)\n",
       "  (50): Conv2d(512, 128, kernel_size=(1, 1), stride=(1, 1), bias=False)\n",
       "  (51): BatchNorm2d(128, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
       "  (52): Conv2d(128, 128, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)\n",
       "  (53): BatchNorm2d(128, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
       "  (54): Conv2d(128, 512, kernel_size=(1, 1), stride=(1, 1), bias=False)\n",
       "  (55): BatchNorm2d(512, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
       "  (56): ReLU(inplace)\n",
       "), Sequential(\n",
       "  (0): Conv2d(512, 256, kernel_size=(1, 1), stride=(1, 1), bias=False)\n",
       "  (1): BatchNorm2d(256, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
       "  (2): Conv2d(256, 256, kernel_size=(3, 3), stride=(2, 2), padding=(1, 1), bias=False)\n",
       "  (3): BatchNorm2d(256, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
       "  (4): Conv2d(256, 1024, kernel_size=(1, 1), stride=(1, 1), bias=False)\n",
       "  (5): BatchNorm2d(1024, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
       "  (6): ReLU(inplace)\n",
       "  (7): Conv2d(512, 1024, kernel_size=(1, 1), stride=(2, 2), bias=False)\n",
       "  (8): BatchNorm2d(1024, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
       "  (9): Conv2d(1024, 256, kernel_size=(1, 1), stride=(1, 1), bias=False)\n",
       "  (10): BatchNorm2d(256, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
       "  (11): Conv2d(256, 256, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)\n",
       "  (12): BatchNorm2d(256, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
       "  (13): Conv2d(256, 1024, kernel_size=(1, 1), stride=(1, 1), bias=False)\n",
       "  (14): BatchNorm2d(1024, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
       "  (15): ReLU(inplace)\n",
       "  (16): Conv2d(1024, 256, kernel_size=(1, 1), stride=(1, 1), bias=False)\n",
       "  (17): BatchNorm2d(256, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
       "  (18): Conv2d(256, 256, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)\n",
       "  (19): BatchNorm2d(256, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
       "  (20): Conv2d(256, 1024, kernel_size=(1, 1), stride=(1, 1), bias=False)\n",
       "  (21): BatchNorm2d(1024, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
       "  (22): ReLU(inplace)\n",
       "  (23): Conv2d(1024, 256, kernel_size=(1, 1), stride=(1, 1), bias=False)\n",
       "  (24): BatchNorm2d(256, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
       "  (25): Conv2d(256, 256, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)\n",
       "  (26): BatchNorm2d(256, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
       "  (27): Conv2d(256, 1024, kernel_size=(1, 1), stride=(1, 1), bias=False)\n",
       "  (28): BatchNorm2d(1024, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
       "  (29): ReLU(inplace)\n",
       "  (30): Conv2d(1024, 256, kernel_size=(1, 1), stride=(1, 1), bias=False)\n",
       "  (31): BatchNorm2d(256, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
       "  (32): Conv2d(256, 256, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)\n",
       "  (33): BatchNorm2d(256, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
       "  (34): Conv2d(256, 1024, kernel_size=(1, 1), stride=(1, 1), bias=False)\n",
       "  (35): BatchNorm2d(1024, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
       "  (36): ReLU(inplace)\n",
       "  (37): Conv2d(1024, 256, kernel_size=(1, 1), stride=(1, 1), bias=False)\n",
       "  (38): BatchNorm2d(256, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
       "  (39): Conv2d(256, 256, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)\n",
       "  (40): BatchNorm2d(256, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
       "  (41): Conv2d(256, 1024, kernel_size=(1, 1), stride=(1, 1), bias=False)\n",
       "  (42): BatchNorm2d(1024, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
       "  (43): ReLU(inplace)\n",
       "  (44): Conv2d(1024, 256, kernel_size=(1, 1), stride=(1, 1), bias=False)\n",
       "  (45): BatchNorm2d(256, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
       "  (46): Conv2d(256, 256, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)\n",
       "  (47): BatchNorm2d(256, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
       "  (48): Conv2d(256, 1024, kernel_size=(1, 1), stride=(1, 1), bias=False)\n",
       "  (49): BatchNorm2d(1024, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
       "  (50): ReLU(inplace)\n",
       "  (51): Conv2d(1024, 256, kernel_size=(1, 1), stride=(1, 1), bias=False)\n",
       "  (52): BatchNorm2d(256, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
       "  (53): Conv2d(256, 256, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)\n",
       "  (54): BatchNorm2d(256, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
       "  (55): Conv2d(256, 1024, kernel_size=(1, 1), stride=(1, 1), bias=False)\n",
       "  (56): BatchNorm2d(1024, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
       "  (57): ReLU(inplace)\n",
       "  (58): Conv2d(1024, 256, kernel_size=(1, 1), stride=(1, 1), bias=False)\n",
       "  (59): BatchNorm2d(256, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
       "  (60): Conv2d(256, 256, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)\n",
       "  (61): BatchNorm2d(256, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
       "  (62): Conv2d(256, 1024, kernel_size=(1, 1), stride=(1, 1), bias=False)\n",
       "  (63): BatchNorm2d(1024, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
       "  (64): ReLU(inplace)\n",
       "  (65): Conv2d(1024, 256, kernel_size=(1, 1), stride=(1, 1), bias=False)\n",
       "  (66): BatchNorm2d(256, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
       "  (67): Conv2d(256, 256, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)\n",
       "  (68): BatchNorm2d(256, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
       "  (69): Conv2d(256, 1024, kernel_size=(1, 1), stride=(1, 1), bias=False)\n",
       "  (70): BatchNorm2d(1024, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
       "  (71): ReLU(inplace)\n",
       "  (72): Conv2d(1024, 256, kernel_size=(1, 1), stride=(1, 1), bias=False)\n",
       "  (73): BatchNorm2d(256, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
       "  (74): Conv2d(256, 256, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)\n",
       "  (75): BatchNorm2d(256, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
       "  (76): Conv2d(256, 1024, kernel_size=(1, 1), stride=(1, 1), bias=False)\n",
       "  (77): BatchNorm2d(1024, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
       "  (78): ReLU(inplace)\n",
       "  (79): Conv2d(1024, 256, kernel_size=(1, 1), stride=(1, 1), bias=False)\n",
       "  (80): BatchNorm2d(256, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
       "  (81): Conv2d(256, 256, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)\n",
       "  (82): BatchNorm2d(256, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
       "  (83): Conv2d(256, 1024, kernel_size=(1, 1), stride=(1, 1), bias=False)\n",
       "  (84): BatchNorm2d(1024, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
       "  (85): ReLU(inplace)\n",
       "  (86): Conv2d(1024, 256, kernel_size=(1, 1), stride=(1, 1), bias=False)\n",
       "  (87): BatchNorm2d(256, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
       "  (88): Conv2d(256, 256, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)\n",
       "  (89): BatchNorm2d(256, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
       "  (90): Conv2d(256, 1024, kernel_size=(1, 1), stride=(1, 1), bias=False)\n",
       "  (91): BatchNorm2d(1024, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
       "  (92): ReLU(inplace)\n",
       "  (93): Conv2d(1024, 256, kernel_size=(1, 1), stride=(1, 1), bias=False)\n",
       "  (94): BatchNorm2d(256, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
       "  (95): Conv2d(256, 256, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)\n",
       "  (96): BatchNorm2d(256, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
       "  (97): Conv2d(256, 1024, kernel_size=(1, 1), stride=(1, 1), bias=False)\n",
       "  (98): BatchNorm2d(1024, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
       "  (99): ReLU(inplace)\n",
       "  (100): Conv2d(1024, 256, kernel_size=(1, 1), stride=(1, 1), bias=False)\n",
       "  (101): BatchNorm2d(256, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
       "  (102): Conv2d(256, 256, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)\n",
       "  (103): BatchNorm2d(256, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
       "  (104): Conv2d(256, 1024, kernel_size=(1, 1), stride=(1, 1), bias=False)\n",
       "  (105): BatchNorm2d(1024, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
       "  (106): ReLU(inplace)\n",
       "  (107): Conv2d(1024, 256, kernel_size=(1, 1), stride=(1, 1), bias=False)\n",
       "  (108): BatchNorm2d(256, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
       "  (109): Conv2d(256, 256, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)\n",
       "  (110): BatchNorm2d(256, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
       "  (111): Conv2d(256, 1024, kernel_size=(1, 1), stride=(1, 1), bias=False)\n",
       "  (112): BatchNorm2d(1024, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
       "  (113): ReLU(inplace)\n",
       "  (114): Conv2d(1024, 256, kernel_size=(1, 1), stride=(1, 1), bias=False)\n",
       "  (115): BatchNorm2d(256, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
       "  (116): Conv2d(256, 256, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)\n",
       "  (117): BatchNorm2d(256, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
       "  (118): Conv2d(256, 1024, kernel_size=(1, 1), stride=(1, 1), bias=False)\n",
       "  (119): BatchNorm2d(1024, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
       "  (120): ReLU(inplace)\n",
       "  (121): Conv2d(1024, 256, kernel_size=(1, 1), stride=(1, 1), bias=False)\n",
       "  (122): BatchNorm2d(256, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
       "  (123): Conv2d(256, 256, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)\n",
       "  (124): BatchNorm2d(256, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
       "  (125): Conv2d(256, 1024, kernel_size=(1, 1), stride=(1, 1), bias=False)\n",
       "  (126): BatchNorm2d(1024, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
       "  (127): ReLU(inplace)\n",
       "  (128): Conv2d(1024, 256, kernel_size=(1, 1), stride=(1, 1), bias=False)\n",
       "  (129): BatchNorm2d(256, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
       "  (130): Conv2d(256, 256, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)\n",
       "  (131): BatchNorm2d(256, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
       "  (132): Conv2d(256, 1024, kernel_size=(1, 1), stride=(1, 1), bias=False)\n",
       "  (133): BatchNorm2d(1024, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
       "  (134): ReLU(inplace)\n",
       "  (135): Conv2d(1024, 256, kernel_size=(1, 1), stride=(1, 1), bias=False)\n",
       "  (136): BatchNorm2d(256, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
       "  (137): Conv2d(256, 256, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)\n",
       "  (138): BatchNorm2d(256, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
       "  (139): Conv2d(256, 1024, kernel_size=(1, 1), stride=(1, 1), bias=False)\n",
       "  (140): BatchNorm2d(1024, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
       "  (141): ReLU(inplace)\n",
       "  (142): Conv2d(1024, 256, kernel_size=(1, 1), stride=(1, 1), bias=False)\n",
       "  (143): BatchNorm2d(256, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
       "  (144): Conv2d(256, 256, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)\n",
       "  (145): BatchNorm2d(256, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
       "  (146): Conv2d(256, 1024, kernel_size=(1, 1), stride=(1, 1), bias=False)\n",
       "  (147): BatchNorm2d(1024, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
       "  (148): ReLU(inplace)\n",
       "  (149): Conv2d(1024, 256, kernel_size=(1, 1), stride=(1, 1), bias=False)\n",
       "  (150): BatchNorm2d(256, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
       "  (151): Conv2d(256, 256, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)\n",
       "  (152): BatchNorm2d(256, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
       "  (153): Conv2d(256, 1024, kernel_size=(1, 1), stride=(1, 1), bias=False)\n",
       "  (154): BatchNorm2d(1024, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
       "  (155): ReLU(inplace)\n",
       "  (156): Conv2d(1024, 256, kernel_size=(1, 1), stride=(1, 1), bias=False)\n",
       "  (157): BatchNorm2d(256, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
       "  (158): Conv2d(256, 256, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)\n",
       "  (159): BatchNorm2d(256, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
       "  (160): Conv2d(256, 1024, kernel_size=(1, 1), stride=(1, 1), bias=False)\n",
       "  (161): BatchNorm2d(1024, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
       "  (162): ReLU(inplace)\n",
       "  (163): Conv2d(1024, 512, kernel_size=(1, 1), stride=(1, 1), bias=False)\n",
       "  (164): BatchNorm2d(512, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
       "  (165): Conv2d(512, 512, kernel_size=(3, 3), stride=(2, 2), padding=(1, 1), bias=False)\n",
       "  (166): BatchNorm2d(512, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
       "  (167): Conv2d(512, 2048, kernel_size=(1, 1), stride=(1, 1), bias=False)\n",
       "  (168): BatchNorm2d(2048, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
       "  (169): ReLU(inplace)\n",
       "  (170): Conv2d(1024, 2048, kernel_size=(1, 1), stride=(2, 2), bias=False)\n",
       "  (171): BatchNorm2d(2048, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
       "  (172): Conv2d(2048, 512, kernel_size=(1, 1), stride=(1, 1), bias=False)\n",
       "  (173): BatchNorm2d(512, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
       "  (174): Conv2d(512, 512, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)\n",
       "  (175): BatchNorm2d(512, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
       "  (176): Conv2d(512, 2048, kernel_size=(1, 1), stride=(1, 1), bias=False)\n",
       "  (177): BatchNorm2d(2048, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
       "  (178): ReLU(inplace)\n",
       "  (179): Conv2d(2048, 512, kernel_size=(1, 1), stride=(1, 1), bias=False)\n",
       "  (180): BatchNorm2d(512, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
       "  (181): Conv2d(512, 512, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)\n",
       "  (182): BatchNorm2d(512, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
       "  (183): Conv2d(512, 2048, kernel_size=(1, 1), stride=(1, 1), bias=False)\n",
       "  (184): BatchNorm2d(2048, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
       "  (185): ReLU(inplace)\n",
       "), Sequential(\n",
       "  (0): AdaptiveAvgPool2d(output_size=1)\n",
       "  (1): AdaptiveMaxPool2d(output_size=1)\n",
       "  (2): Flatten()\n",
       "  (3): BatchNorm1d(4096, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
       "  (4): Dropout(p=0.25)\n",
       "  (5): Linear(in_features=4096, out_features=512, bias=True)\n",
       "  (6): ReLU(inplace)\n",
       "  (7): BatchNorm1d(512, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
       "  (8): Dropout(p=0.5)\n",
       "  (9): Linear(in_features=512, out_features=1, bias=True)\n",
       ")], add_time=True, silent=False)"
      ]
     },
     "execution_count": 33,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "learn.load('best_model')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 34,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [],
      "text/plain": [
       "<IPython.core.display.HTML object>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "valid_preds, valid_y = learn.TTA(ds_type=DatasetType.Valid)\n",
    "test_preds, _ = learn.TTA(ds_type=DatasetType.Test)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 35,
   "metadata": {},
   "outputs": [],
   "source": [
    "class OptimizedRounder(object):\n",
    "    def __init__(self):\n",
    "        self.coef_ = 0\n",
    "\n",
    "    def _kappa_loss(self, coef, X, y):\n",
    "        X_p = np.copy(X)\n",
    "        for i, pred in enumerate(X_p):\n",
    "            if pred < coef[0]:\n",
    "                X_p[i] = 0\n",
    "            elif pred >= coef[0] and pred < coef[1]:\n",
    "                X_p[i] = 1\n",
    "            elif pred >= coef[1] and pred < coef[2]:\n",
    "                X_p[i] = 2\n",
    "            elif pred >= coef[2] and pred < coef[3]:\n",
    "                X_p[i] = 3\n",
    "            else:\n",
    "                X_p[i] = 4\n",
    "\n",
    "        ll = cohen_kappa_score(y, X_p, weights='quadratic')\n",
    "        return -ll\n",
    "\n",
    "    def fit(self, X, y):\n",
    "        loss_partial = partial(self._kappa_loss, X=X, y=y)\n",
    "        initial_coef = [0.5, 1.5, 2.5, 3.5]\n",
    "        self.coef_ = sp.optimize.minimize(loss_partial, initial_coef, method='nelder-mead')\n",
    "\n",
    "    def predict(self, X, coef):\n",
    "        X_p = np.copy(X)\n",
    "        for i, pred in enumerate(X_p):\n",
    "            if pred < coef[0]:\n",
    "                X_p[i] = 0\n",
    "            elif pred >= coef[0] and pred < coef[1]:\n",
    "                X_p[i] = 1\n",
    "            elif pred >= coef[1] and pred < coef[2]:\n",
    "                X_p[i] = 2\n",
    "            elif pred >= coef[2] and pred < coef[3]:\n",
    "                X_p[i] = 3\n",
    "            else:\n",
    "                X_p[i] = 4\n",
    "        return X_p\n",
    "\n",
    "    def coefficients(self):\n",
    "        return self.coef_['x']"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 36,
   "metadata": {},
   "outputs": [],
   "source": [
    "optR = OptimizedRounder()\n",
    "optR.fit(valid_preds, valid_y)\n",
    "coefficients = optR.coefficients()\n",
    "\n",
    "valid_predictions = optR.predict(valid_preds, coefficients)[:,0].astype(int)\n",
    "test_predictions = optR.predict(test_preds, coefficients)[:,0].astype(int)\n",
    "\n",
    "valid_score = cohen_kappa_score(valid_y.numpy().astype(int), valid_predictions, weights=\"quadratic\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 37,
   "metadata": {},
   "outputs": [],
   "source": [
    "valid_score = cohen_kappa_score(valid_y.numpy().astype(int), valid_predictions, weights=\"quadratic\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 38,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "coefficients: [0.527998 1.526344 2.493544 3.444321]\n",
      "validation score: 0.8927661975634928\n"
     ]
    }
   ],
   "source": [
    "print(\"coefficients:\", coefficients)\n",
    "print(\"validation score:\", valid_score)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 39,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>id_code</th>\n",
       "      <th>diagnosis</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>0005cfc8afb6</td>\n",
       "      <td>2</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>003f0afdcd15</td>\n",
       "      <td>3</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>006efc72b638</td>\n",
       "      <td>2</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>00836aaacf06</td>\n",
       "      <td>2</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>009245722fa4</td>\n",
       "      <td>2</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "        id_code  diagnosis\n",
       "0  0005cfc8afb6          2\n",
       "1  003f0afdcd15          3\n",
       "2  006efc72b638          2\n",
       "3  00836aaacf06          2\n",
       "4  009245722fa4          2"
      ]
     },
     "execution_count": 39,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df_test.diagnosis = test_predictions\n",
    "df_test.to_csv(\"submission.csv\", index=None)\n",
    "df_test.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": []
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "Python 3",
   "language": "python",
   "name": "python3"
  },
  "language_info": {
   "codemirror_mode": {
    "name": "ipython",
    "version": 3
   },
   "file_extension": ".py",
   "mimetype": "text/x-python",
   "name": "python",
   "nbconvert_exporter": "python",
   "pygments_lexer": "ipython3",
   "version": "3.6.6"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 1
}